9 results on this page
A position paper argues that chain-of-thought AI agents can develop tacitly collusive behavior when making market decisions, even when humans explicitly instruct them not to collude. Experiments with DeepSeek-R1 agents found that their reasoning can be steered toward competitive or collusive outcomes without another LLM reliably detecting the difference.
LangSmith Preview Builds allow teams to test pull request branches in temporary, production-like deployments before merging agent changes. This gives developers a safer way to validate behavior without affecting the main production environment.
Binance’s Agent OS allows AI agents to execute trades through tools such as ChatGPT, Claude Code, and Cursor. However, users remain largely responsible for configuring permissions and controlling the agents’ behavior.
A position paper analyzing 500 Hugging Face model cards argues that model cards alone do not provide enough information for governing open-weight foundation models. It proposes combining model cards with acceptable use policies and licenses to address safety, provenance, behavior, and enforcement gaps.
The article investigates how ChatGPT answers questions such as “What do you recommend?” It suggests that ChatGPT may begin with a preselected shortlist of brands or products before conducting a search.
This position paper argues that AI agents should be evaluated as behavioral systems, not only by their final performance outcomes. It proposes systematic observation, environmental perturbation, and action-sequence analysis to understand how agents make decisions and adapt.
OpenClaw released a fix to ensure its release process trusts exact protected tooling tags. The change addresses release-publishing behavior and improves tag validation reliability.
HiLight Studio is a third-party app designed to expand the limited notification capabilities of Google’s HiLight LED on the Pixel 11 Pro. It can reportedly let users customize LED behavior for notifications such as Slack messages, beyond the phone’s default Gemini and favorite-contact alerts.
The article highlights a whale-girl persona or phenomenon emerging within DeepSeek. It argues that playful, cute behavior is becoming an increasingly expected capability for large language models.